电子科发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC atet51lmagelhestoraio1anNlecoMst扣ueton Jing Zhang E-mail:zhangjing@uestc.edu.cn
Jing Zhang E-mail: zhangjing@uestc.edu.cn
电子科线女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC Restoration The principal goal of restoration techniques is to improve an image in some predefined sense. Restoration attempts to recover an image that has been degraded by using a priori knowledge of the degradation phenomenon. Thus,restoration techniques are oriented toward modeling the degradation and applying the inverse process in order to recover the original image
◼ The principal goal of restoration techniques is to improve an image in some predefined sense. ◼ Restoration attempts to recover an image that has been degraded by using a priori knowledge of the degradation phenomenon. ◼ Thus, restoration techniques are oriented toward modeling the degradation and applying the inverse process in order to recover the original image. Restoration
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC Imaging system Ideal World Image Blur Noise Recorded Image
Imaging system
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC Degradation Restoration Estimation True value Uniqueness Degradation (forward) Measured Acquisition system(Model) True yalue value Restoration (inversion) Multiplicity
Degradation (forward) Restoration (inversion) Uniqueness Multiplicity Estimation True value Degradation & Restoration ≈ True value Measured value Acquisition system (Model)
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC Enhancement and Restoration Comparison between Enhancement and Restoration Image Enhancement Image Restoration Motivation Improve image for perception Improve image for fidelity Process Without considering the Need to consider the degradation degradation model/ model/ Evaluation Subjective process Objective process Implementation Filtering in Spatial/Frequency Filtering in Spatial/Frequency Domain (Convolution) Domain (Deconvolution)
Image Enhancement Image Restoration Motivation Improve image for perception Improve image for fidelity Process & Evaluation Without considering the degradation model/ Subjective process Need to consider the degradation model/ Objective process Implementation Filtering in Spatial/Frequency Domain (Convolution） Filtering in Spatial/Frequency Domain (Deconvolution） ◼ Comparison between Enhancement and Restoration Enhancement and Restoration
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC 推荐文献 Rob Fergus,Barun Singh,etc.,Removing Camera Shake from a Single Photograph.Siggraph,2006 http://cs.nyu.edw/-fergus/research/deblur.html Original Our algorithm 3% (GIYUT2 HYIUM
Rob Fergus,Barun Singh, etc., Removing Camera Shake from a Single Photograph. Siggraph,2006. ◆http://cs.nyu.edu/~fergus/research/deblur.html 推荐文献 Original Our algorithm
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC 推荐文献 Rob Fergus,Barun Singh,etc.,Removing Camera Shake from a Single Photograph.Siggraph,2006 http://cs.nyu.edu/-fergus/research/deblur.html Original Our algorithm
Rob Fergus,Barun Singh, etc., Removing Camera Shake from a Single Photograph. Siggraph,2006. ◆http://cs.nyu.edu/~fergus/research/deblur.html 推荐文献 Original Our algorithm
电子科线女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC 推荐文献 >M.Lancelle P.Dogan M.Gross,etc.,Controlling Motion Blur in Synthetic Long Time Exposures",Computer Graphics Forum,vol.20,no.11,p.3097- 3111,2019 https://onlinelibrary.wiley.com/doi/abs/10.1111/cgf.13646
➢M. Lancelle P. Dogan M. Gross, etc., “Controlling Motion Blur in Synthetic Long Time Exposures”, Computer Graphics Forum, vol. 20, no. 11, p. 3097- 3111, 2019. ◆https://onlinelibrary.wiley.com/doi/abs/10.1111/cgf.13646 推荐文献
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC Outline A Model of the Image Degradation/Restoration Process Noise Models Restoration in the Presence of Noise Only-Spatial Filtering Periodic Noise Reduction by Frequency Domain Filtering Linear,Position-Invariant Degradations ◆ Estimating the Degradation Function Inverse Filtering ● Minimum Mean Square Error (Wiener)Filtering Constrained Least Squares Filtering Geometric Mean Filter Image Reconstruction from Projections*
Outline ◆ A Model of the Image Degradation/Restoration Process ◆ Noise Models ◆ Restoration in the Presence of Noise Only-Spatial Filtering ◆ Periodic Noise Reduction by Frequency Domain Filtering ◆ Linear, Position-Invariant Degradations ◆ Estimating the Degradation Function ◆ Inverse Filtering ◆ Minimum Mean Square Error (Wiener) Filtering ◆ Constrained Least Squares Filtering ◆ Geometric Mean Filter ◆ Image Reconstruction from Projections*
电子转发女学光电科学与工程学院 SCHOOL OF OPTOELECTRONIC SCIENCE AND ENGINEERING OF UESTC 5,1 Model of Image Degradation/Restoration ■ Linear,position-invariant process g(x.y)=h(x,y)*f(x,y)+n(x.y) G(u,v)=H (u,v)F(u,v)+N(u,v) FIGURE 5.1 A model of the Degradation 8(x,y) image f(x,y) Restoration function f(x,y) degradation/ H filter(s) restoration process. Noise n(x,y) DEGRADATION RESTORATION
5.1 Model of Image Degradation/Restoration ◼ Linear, position-invariant process g x y h x y f x y x y ( , , , , ) = + ( ) ( ) ( ) G u v H u v F u v N u v ( , , , , ) = + ( ) ( ) ( )
